CN111275121A - Medical image processing method and device and electronic equipment - Google Patents
Medical image processing method and device and electronic equipment Download PDFInfo
- Publication number
- CN111275121A CN111275121A CN202010076247.7A CN202010076247A CN111275121A CN 111275121 A CN111275121 A CN 111275121A CN 202010076247 A CN202010076247 A CN 202010076247A CN 111275121 A CN111275121 A CN 111275121A
- Authority
- CN
- China
- Prior art keywords
- medical image
- lesion
- image picture
- network model
- hierarchical
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30096—Tumor; Lesion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/03—Recognition of patterns in medical or anatomical images
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Medical Informatics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Probability & Statistics with Applications (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
- Apparatus For Radiation Diagnosis (AREA)
- Image Analysis (AREA)
Abstract
The application discloses a medical image processing method and device and electronic equipment, and relates to the field of image processing in the field of computers. The specific implementation scheme is as follows: a medical image processing method, comprising: acquiring a medical image picture; inputting the medical image picture into a hierarchical network model obtained by pre-training, calculating the medical image picture by the hierarchical network model to output a hierarchical array comprising C-1 elements, wherein the medical image picture corresponds to C lesion grades, the C lesion grades comprise 1 lesion grade without lesions and C-1 lesion grade with lesions, the kth element of the hierarchical array is the probability that the lesion grade corresponding to the medical image picture is greater than or equal to the kth grade, and k is greater than or equal to 1 and is less than or equal to C-1. The medical image processing method, the medical image processing device and the electronic equipment can solve the problem that the method in the prior art is poor in medical image processing effect.
Description
Technical Field
The present application relates to the field of image processing in the field of computers, and in particular, to a medical image processing method and apparatus, and an electronic device.
Background
With the continuous development and progress of medical imaging technology and computer technology, medical image analysis has become an indispensable tool and technical means in medical research, clinical disease diagnosis and treatment, wherein, by processing medical image pictures and obtaining corresponding processing results, reference can be provided for subsequent treatment. In the prior art, a medical image is generally input into a neural network model, and a focus region in the medical image is identified and calculated by the neural network model to output a corresponding processing result.
Disclosure of Invention
The application provides a medical image processing method, a medical image processing device and electronic equipment, and aims to solve the problem that the method in the prior art is poor in medical image processing effect.
In a first aspect, the present application provides a medical image processing method, including:
acquiring a medical image picture;
inputting the medical image picture into a hierarchical network model obtained by pre-training, calculating the medical image picture by the hierarchical network model, and outputting a hierarchical array comprising C-1 elements, wherein the medical image picture corresponds to C lesion levels, the C lesion levels comprise 1 lesion level without lesions and C-1 lesion level with lesions, the kth element of the hierarchical array is the probability that the lesion level corresponding to the medical image picture is greater than or equal to the kth level, and k is greater than or equal to 1 and is less than or equal to C-1.
Therefore, when the medical image picture is processed, the probability that the lesion grade corresponding to the medical image picture is larger than or equal to each lesion grade is calculated respectively, so that even if one probability value has an error, the lesion grade corresponding to the medical image picture can be predicted by combining other probability values, the problem that the medical image picture processing effect is poor by using the method in the prior art is solved, and the reliability of the medical image picture processing result is improved.
Optionally, before the medical image picture is acquired, the method further comprises:
constructing a convolutional neural network model;
and inputting training data to the convolutional neural network model, and training the convolutional neural network model according to the training data to obtain the hierarchical network model.
In the embodiment, the hierarchical network model is constructed and trained by adopting the convolutional neural network technology, so that the film reading efficiency can be improved.
Optionally, the training data includes a sample image picture and a target array corresponding to the sample image picture, and the target array includes C-1 elements;
when the lesion grade corresponding to the sample image picture is the lesion grade without lesions, the values of C-1 elements in the target array are all 0;
and when the lesion grade corresponding to the sample image picture is the kth grade, the values of the first k elements of the target array are all 1, and the values of the (k + 1) th to the (C-1) th elements of the target array are all 0.
In the embodiment, the method is adopted to train the hierarchical network model, so that the reliability of the trained hierarchical network model can be effectively improved.
Optionally, before the medical image picture is input into a hierarchical network model trained in advance, the method further includes:
respectively carrying out normalization processing and filtering processing on the medical image picture to obtain a preprocessed picture;
the step of inputting the medical image picture into a hierarchical network model obtained by pre-training, and calculating the medical image picture by the hierarchical network model to output a hierarchical array comprising C-1 elements comprises:
inputting the preprocessed picture into a hierarchical network model obtained by pre-training, and calculating the preprocessed picture by the hierarchical network model to output a hierarchical array comprising C-1 elements.
In the embodiment, the medical image picture input into the hierarchical network model is preprocessed, so that the image picture input into the hierarchical network model has a clear focus area, and the calculation accuracy of the hierarchical network model can be improved.
Optionally, after the medical image is input into a hierarchical network model obtained by pre-training, and the medical image is calculated by the hierarchical network model to output a hierarchical array including C-1 elements, the method further includes:
and determining the lesion grade corresponding to the medical image picture according to the grading array.
In the embodiment, after the grading array is obtained, the lesion grade corresponding to the medical image picture is further determined according to the grading array, so that the process of reading the picture is further simplified, and the efficiency of reading the picture can be further improved.
In a second aspect, the present application provides a medical image processing apparatus, including:
the acquisition module is used for acquiring a medical image picture;
the calculation module is used for inputting the medical image picture into a hierarchical network model obtained by pre-training, calculating the medical image picture by the hierarchical network model and outputting a hierarchical array comprising C-1 elements, wherein the medical image picture corresponds to C lesion levels, the C lesion levels comprise 1 lesion level without lesion and C-1 lesion level with lesion, the kth element of the hierarchical array is the probability that the lesion level corresponding to the medical image picture is greater than or equal to the kth level, and k is greater than or equal to 1 and is less than or equal to C-1.
Optionally, the apparatus further comprises:
the construction module is used for constructing a convolutional neural network model before the medical image picture is acquired;
and the training module is used for inputting training data to the convolutional neural network model and training the convolutional neural network model according to the training data to obtain the hierarchical network model.
Optionally, the training data includes a sample image picture and a target array corresponding to the sample image picture, and the target array includes C-1 elements;
when the lesion grade corresponding to the sample image picture is the lesion grade without lesions, the values of C-1 elements in the target array are all 0;
and when the lesion grade corresponding to the sample image picture is the kth grade, the values of the first k elements of the target array are all 1, and the values of the (k + 1) th to the (C-1) th elements of the target array are all 0.
In a third aspect, the present application provides an electronic device, comprising:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores instructions executable by the at least one processor, the instructions being executable by the at least one processor to enable the at least one processor to perform the medical image processing method provided herein.
In a fourth aspect, the present application provides a non-transitory computer-readable storage medium storing computer instructions for causing a computer to execute the medical image processing method provided by the present application.
One embodiment in the above application has the following advantages or benefits: when the medical image picture is processed, the probability that the lesion grade corresponding to the medical image picture is larger than or equal to each lesion grade is calculated respectively, so that the lesion grade corresponding to the medical image picture can be predicted by combining other probability values even if errors exist in part of the calculated probability values, the problem that the medical image picture processing effect is poor by using the method in the prior art is solved, and the reliability of the medical image picture processing result is improved.
Other effects of the above-described alternative will be described below with reference to specific embodiments.
Drawings
The drawings are included to provide a better understanding of the present solution and are not intended to limit the present application. Wherein:
fig. 1 is a flowchart of a medical image processing method provided in an embodiment of the present application;
FIG. 2 is a flow diagram of a hierarchical network model training process in an embodiment of the present application;
fig. 3 is a second flowchart of a medical image processing method provided in an embodiment of the present application;
FIG. 4 is a flowchart of determining a grade of diabetic retinopathy in an embodiment of the present application;
FIG. 5 is a schematic structural diagram of a medical image processing apparatus according to an embodiment of the present application;
FIG. 6 is a second schematic structural diagram of a medical image processing apparatus according to an embodiment of the present application;
fig. 7 is a block diagram of an electronic device for implementing processing of medical images according to an embodiment of the present application.
Detailed Description
The following description of the exemplary embodiments of the present application, taken in conjunction with the accompanying drawings, includes various details of the embodiments of the application for the understanding of the same, which are to be considered exemplary only. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present application. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.
Referring to fig. 1, fig. 1 is a medical image processing method provided in the present application, including:
and S101, acquiring a medical image picture.
In the present application, the medical image may be an image acquired for analyzing whether there is a certain lesion in a human or animal body, for example, a fundus image for determining diabetic retinopathy, a breast image for determining whether there is tuberculosis, or the like, which is not limited thereto.
The medical image picture can be acquired by medical imaging equipment, for example, a fundus image picture acquired by a Baidu AI fundus screening all-in-one machine.
It should be understood that the medical image picture may be an image picture showing that no lesion exists in the human or animal body, or an image picture showing that a lesion exists in the human or animal body. The image picture showing the existence of the lesion in the human or animal body may include a diagnostic mark region (or lesion region) for diagnosing the degree of the lesion and other regions without diagnostic marks, and the existence of the lesion may be determined by determining the existence of the diagnostic mark region, or the severity of the lesion may be determined by analyzing the diagnostic mark region.
Step S102, inputting the medical image picture into a hierarchical network model obtained through pre-training, calculating the medical image picture through the hierarchical network model, and outputting a hierarchical array comprising C-1 elements, wherein the medical image picture corresponds to C lesion levels, the C lesion levels comprise 1 lesion level without lesions and C-1 lesion level with lesions, the kth element of the hierarchical array is the probability that the lesion level corresponding to the medical image picture is greater than or equal to the kth level, and k is greater than or equal to 1 and is less than or equal to C-1.
The hierarchical Network model may adopt any classical classification Network, such as an inclusion structure (inclusion Construction), a Residual Network (ResNet), a dense Connected Convolutional Network (densneet), and the like, or may be a self-constructed classification Network.
Because the hierarchical network model needs to output an array formed by a plurality of probability values, and the value range of each probability value is between 0 and 1, the hierarchical network model can be a hierarchical network model with a sigmoid activation function, namely, an activation function layer is additionally arranged at the output end of the hierarchical network model, the output of the hierarchical network model is used as the input of the activation function layer, and the data processed by the hierarchical network model is mapped between 0 and 1 by the activation function layer.
Each element in the hierarchical array is a group of data which are arranged in order, wherein each element corresponds to a lesion grade, for example, the number of lesion levels C is equal to 5, and the 5 lesion levels include 1 lesion level without lesion (level 0) and 4 lesion levels with lesion, the 4 lesion levels with lesion classify the lesion levels into levels 1, 2, 3, and 4 according to the severity of lesion, at which time, the hierarchical array includes 4 elements, for example, the hierarchical array may be [ a, b, c, d ], where a corresponds to a probability that the lesion level of the medical image is greater than or equal to level 1, b corresponds to a probability that the lesion level of the medical image is greater than or equal to level 2, c corresponds to a probability that the lesion level of the medical image is greater than or equal to level 3, and d corresponds to a probability that the lesion level of the medical image is greater than or equal to level 4. If the lesion level corresponding to the medical image is known to be level 3, since the lesion level is necessarily greater than or equal to level 1, level 2, and level 3 but not necessarily greater than or equal to level 4 when the lesion level is level 3, the probability values represented by a, b, and c are all 100%, and the probability value represented by d is 0%, so the rank array can be represented as [1,1,1,0 ]. Similarly, when the lesion level corresponding to the known medical image is level 1, the hierarchical array may be represented as [1,0,0,0], when the lesion level corresponding to the known medical image is level 2, the hierarchical array may be represented as [1,1,0,0], when the lesion level corresponding to the known medical image is level 4, the hierarchical array may be represented as [1,1,1,1], and when the lesion level corresponding to the medical image is level 0, since no lesion exists at this time, the lesion level corresponding to the medical image is inevitably not greater than or equal to level 1, level 2, level 3, and level 4, and therefore, the probability values represented by a, b, c, d are all 0%, and at this time, the hierarchical array may be represented as [0,0,0,0 ].
When the calculation is performed through the hierarchical network model, the hierarchical network model may calculate C-1 times respectively, that is, each probability value is calculated independently, so that even if there is a deviation between a certain calculation result and an actual result due to an error of the hierarchical network model or other reasons, the error between the calculated result and the actual result is not too large, for example, a lesion level corresponding to a certain medical image picture is level 4, while when the hierarchical network model is calculated, a failure occurs in one calculation, and an output result is: and [1,1,1,0], [1,0,1,1] or [1,1,0.6,1], etc., when the medical staff looks at the output result, the medical staff can basically judge that the lesion grade corresponding to the medical image picture is possibly above the 3 rd grade or even above the 4 th grade, or find that the calculation result has a problem, and perform remedial measures such as recalculation or manual reading on the calculation result, etc., thereby ensuring the reliability of the calculation result.
In the prior art, a hierarchical network model usually outputs a certain type or a certain level, and when the hierarchical network model has a fault or an output result is inaccurate due to other reasons, a problem that an error between the output result and an actual result is very large may be caused, for example, a lesion level corresponding to a certain medical image picture is level 4, and an output result is level 0 due to an error of the hierarchical network model or other reasons, at this time, if a medical staff excessively depends on a hierarchical prediction result, an error between the prediction result and the actual result is very large, and further, an optimal treatment opportunity is delayed, and the like.
Compared with the prior art, the medical image processing method provided by the application has the advantages that the incidence relation between the input medical image picture and each lesion grade is considered, and the probability that the lesion grade corresponding to the medical image picture is greater than or equal to each lesion grade is output, so that the reliability of the calculation result can be ensured, and the accuracy of lesion grade prediction is improved.
Optionally, referring to fig. 2, before acquiring the medical image picture, the method further includes:
s201, constructing a convolutional neural network model;
step S202, inputting training data to the convolutional neural network model, and training the convolutional neural network model according to the training data to obtain the hierarchical network model.
Specifically, the hierarchical network model is constructed and trained by adopting a convolutional neural network technology, the hierarchical network model can directly calculate the medical image picture and output a hierarchical array comprising C-1 elements, and each element in the hierarchical array respectively represents the probability that the lesion grade corresponding to the medical image picture is greater than or equal to each lesion grade.
Optionally, the training data may include a sample image picture and a target array corresponding to the sample image picture, where the target array includes C-1 elements;
when the lesion grade corresponding to the sample image picture is the lesion grade without lesions, the values of C-1 elements in the target array are all 0;
and when the lesion grade corresponding to the sample image picture is the kth grade, the values of the first k elements of the target array are all 1, and the values of the (k + 1) th to the (C-1) th elements of the target array are all 0.
The sample image picture may be a representative or general image picture acquired in medical practice, and the lesion level corresponding to the sample image picture has been determined through medical practice, and it is determined that the sample image picture is matched with the corresponding lesion level, so that the reliability of the trained hierarchical network model can be improved. As can be seen from the above discussion, when the lesion level is level 0, the target array may be [0,0,0,0], when the lesion level is level 1, the target array may be [1,0,0,0], and so on, and the corresponding relationship between the sample image picture and the target array may be established through the lesion level.
For example, when the number C of lesion levels is equal to 5, the corresponding relationship between each sample image picture and the target array can be established according to the following rules:
during specific training, the gradient descent method is used for adjusting the model parameters according to the difference between the model output array and the target array so as to reduce the difference between the model output array and the target array, and when the difference between the model output array and the target array is reduced to be within an allowable error range, the training of the grading network model can be determined to be completed.
In addition, the sample image picture may be a sample image picture obtained through preprocessing, where the preprocessing may be performed by performing normalization, gaussian filtering, median filtering, and the like on the medical image picture, where the normalization is to normalize the sample image pictures with different brightness levels to the same range. The gaussian filtering process and the median filtering process are used to highlight the physiological structures and the lesion areas in the medical image picture.
Optionally, before the medical image picture is input into a hierarchical network model trained in advance, the method further includes:
respectively carrying out normalization processing and filtering processing on the medical image picture to obtain a preprocessed picture;
the step of inputting the medical image picture into a hierarchical network model obtained by pre-training, and calculating the medical image picture by the hierarchical network model to output a hierarchical array comprising C-1 elements comprises:
inputting the preprocessed picture into a hierarchical network model obtained by pre-training, and calculating the preprocessed picture by the hierarchical network model to output a hierarchical array comprising C-1 elements.
Specifically, the normalization processing may be a Z-Score (Z-Score) normalization processing method, and the filtering processing may be a filtering processing manner such as gaussian filtering processing or median filtering processing, and the image picture input into the hierarchical network model has a clear lesion area by preprocessing the medical image picture input into the hierarchical network model, so that the accuracy of the calculation of the hierarchical network model can be improved.
After the hierarchical array is obtained through the hierarchical network model, the hierarchical array can be directly provided for medical staff who specially process medical image pictures, and the medical staff determines the lesion grade corresponding to the medical image pictures by combining the probability that the lesion grade corresponding to the medical image pictures is greater than or equal to each lesion grade. In addition, the data processing may be further performed on the ranking array, and the lesion level corresponding to the ranking array may be output in consideration of each probability value in the ranking array.
Specifically, in order to implement further data processing on the hierarchical array to output the lesion level corresponding to the hierarchical array, please refer to fig. 3, which is another medical image processing method provided by the present application, the method includes the following steps:
step 301, acquiring a medical image picture;
step 302, inputting the medical image picture into a hierarchical network model obtained by pre-training, calculating the medical image picture by the hierarchical network model, and outputting a hierarchical array comprising C-1 elements, wherein the medical image picture corresponds to C lesion levels, the C lesion levels comprise 1 lesion level without lesion and C-1 lesion level with lesion, the kth element of the hierarchical array is the probability that the lesion level corresponding to the medical image picture is greater than or equal to the kth level, and k is greater than or equal to 1 and is less than or equal to C-1;
and 303, determining the lesion grade corresponding to the medical image picture according to the grading array.
Step 301 and step 302 correspond to step 101 and step 102, respectively, the specific implementation manner is the same as that in the above embodiment, and the same beneficial effects can be achieved.
In the embodiment of the application, after the hierarchical array is obtained in step 302, the lesion grade corresponding to the medical image picture is further determined according to the hierarchical array, so that the process of reading the picture is further simplified, and the efficiency of reading the picture can be further improved.
Optionally, the determining, according to the hierarchical array, a lesion level corresponding to the medical image picture includes:
calculating the sum of C-1 elements of the hierarchical array;
and inquiring the lesion grade corresponding to the medical image picture in a preset database by taking the sum of the elements as an inquiry condition, wherein the preset database stores the corresponding relation between the lesion grade and the sum of each element.
In the prior art, a hierarchical network model usually outputs a certain type or a certain level, and specifically, in the prior art, a one-to-one correspondence relationship between a sample picture and a lesion level is usually established, and when a lesion level of a certain medical image picture is determined by the hierarchical network model, a sample image picture closest to the received medical image picture is usually queried in a sample library, and the lesion level corresponding to the sample image picture is taken as the lesion level of the medical image picture. Thus, when the hierarchical network model has a fault or the output result is inaccurate due to other reasons, the problem that the error between the output result and the actual result is very large can be caused.
In the embodiment of the application, the lesion grade corresponding to the medical image picture is inquired by calculating the sum of each element of the hierarchical array and inquiring by taking the sum of each element as an inquiry condition, so that the incidence relation among the lesion grades can be fully considered to ensure the reliability of the calculation result.
Optionally, in order to fully consider the association relationship between the respective lesion levels to ensure the reliability of the calculation result, before the acquiring the medical image picture, the method further includes:
c-1 division points are selected between the intervals [0, C ], and the intervals [0, C ] are divided to form C sub-intervals;
establishing a corresponding relation between the C lesion grades and the C sub-intervals;
the querying, in a preset database, a lesion level corresponding to the medical image picture with the sum of the elements as a query condition includes:
querying in the C sub-intervals by taking the sum of the elements as a query condition to determine a target sub-interval containing the sum of the elements;
and taking the lesion grade corresponding to the target subinterval as the lesion grade corresponding to the medical image picture.
Wherein, the selecting C-1 division points between the intervals [0, C ] to divide the intervals [0, C ] comprises: and selecting [ a + i ] as a division point, and dividing the interval [0, C ], wherein a is more than 0 and less than 1, i is 0, … and C-2. Optimization algorithms suitable for the non-derivable functions, such as a downhill simplex method (Nelder-Mead), are maximized with a position-sensitive classification precision index (Kappa) as an optimization target, so that the value of a is reduced under the condition that the classification requirements are met, the span of an interval corresponding to a lesion grade with a lesion is increased, the lesion grade corresponding to an image picture is more accurately mapped to a corresponding sub-interval, and the accuracy of lesion grade prediction is further improved.
Specifically, the following explains the determination of the lesion levels corresponding to the medical image picture according to the hierarchical array, taking the number of lesion levels as 5, and taking a as 0.5 as an example, since [ a + i ] is selected as the division point, the division points in the embodiment of the present application can be determined as 0.5, 1.5, 2.5, and 3.5, respectively, so as to divide the interval [0, 4] into the following subintervals [0, 0.5), [0.5, 1.5), [1.5, 2.5), [2.5, 3.5), and [3.5, 4 ]. Wherein, the lesion grade corresponding to the subinterval [0, 0.5) is level 0, that is, when the sum of all elements is less than 0.5, it may be determined that the lesion grade corresponding to the medical image is the lesion grade without lesions, similarly, the lesion grade corresponding to the subinterval [0.5, 1.5) is level 1, the lesion grade corresponding to the subinterval [1.5, 2.5) is level 2, the lesion grade corresponding to the subinterval [2.5, 3.5) is level 3, and the lesion grade corresponding to the subinterval [3.5, 4] is level 4, thereby establishing the corresponding relationship between the probability represented by each element and each lesion grade.
Referring to the following table, the following describes a specific process for determining the grade of a lesion corresponding to a medical image picture according to the sum of elements, in an embodiment:
as can be seen from the above table, when the sum of each element is 0.4, since the sum is included in the subinterval [0, 0.5), the corresponding lesion level is level 0, where the hierarchical array with the sum of the elements being 0.4 may be [0.4,0,0,0], [0.1,0.2,0.1,0], [0,0, 0.4], and the like. For example, when the hierarchical array is [0,0,0,0.4], if according to the conventional method, since the probability that the medical image picture belongs to the 4 th level is calculated according to the hierarchical network model, the conventional method may correspond the corresponding lesion level to the 4 th level, whereas according to the method provided by the embodiment of the present application, since the probability that the medical image picture corresponds to the lesion level greater than or equal to the 1 st level, the 2 nd level, and the 3 rd level is 0, it can be seen that the lesion level corresponding to the medical image picture most likely corresponds to the lesion level that should be the 0 th level rather than the 4 th level, and the calculated probability greater than or equal to the 4 th level is likely to be caused by the existence of a fault in the hierarchical network model or due to other reasons. Similarly, as can be seen from the above table, when the sum of the elements is 1.4, since it is included in the subinterval [0.5, 1.5), the lesion level can be determined to be level 1; when the sum of the elements is 1.6, since it is included in the subinterval [1.5, 2.5), it is determined that the lesion level is level 2; when the sum of the elements is 2.6, since it is included in the subinterval [2.5, 3.5), it is determined that the lesion level is level 3; when the sum of the elements is 3.6, since it is included in the sub-interval [3.5, 4], the lesion level thereof can be determined to be level 4.
Specifically, the probability that the lesion level corresponding to the medical image picture is greater than or equal to each lesion level is calculated respectively in the embodiment of the application, so that even if errors exist in part of the calculated probability values, the lesion level corresponding to the medical image picture can be predicted by combining other probability values, and therefore the problem that the medical image picture processing effect is poor by using the method in the prior art is solved, and the reliability of the medical image picture processing result is improved.
Referring to fig. 4, fig. 4 is a flowchart illustrating the method applied to determine the grade of diabetic retinopathy, which specifically includes the following steps: acquiring an eye fundus picture, preprocessing the eye fundus picture, inputting the preprocessed picture into a pre-selected trained hierarchical network model, calculating by the hierarchical network model and outputting a hierarchical array, calculating the sum of each element in the hierarchical array, and determining the lesion grade corresponding to the eye fundus picture according to the calculated sum of each element and a preset mapping relation. Meanwhile, as can be seen from fig. 4, the hierarchical network model is obtained by inputting training data into the constructed convolutional neural network model and performing training.
Referring to fig. 5, fig. 5 is a diagram illustrating a medical image processing apparatus 500 according to an embodiment of the present application, including:
an obtaining module 501, configured to obtain a medical image picture;
a calculating module 502, configured to input the medical image picture into a pre-trained hierarchical network model, and calculate the medical image picture by using the hierarchical network model to output a hierarchical array including C-1 elements, where the medical image picture corresponds to C lesion levels, the C lesion levels include 1 lesion level without a lesion and C-1 lesion levels with a lesion, a kth element of the hierarchical array is a probability that the lesion level corresponding to the medical image picture is greater than or equal to the kth level, and k is greater than or equal to 1 and is less than or equal to C-1.
Optionally, the medical image processing apparatus 500 further includes:
a building module 503, configured to build a convolutional neural network model before acquiring the medical image picture;
a training module 504, configured to input training data to the convolutional neural network model, and train the convolutional neural network model according to the training data to obtain the hierarchical network model.
Optionally, the training data includes a sample image picture and a target array corresponding to the sample image picture, and the target array includes C-1 elements;
when the lesion grade corresponding to the sample image picture is the lesion grade without lesions, the values of C-1 elements in the target array are all 0;
and when the lesion grade corresponding to the sample image picture is the kth grade, the values of the first k elements of the target array are all 1, and the values of the (k + 1) th to the (C-1) th elements of the target array are all 0.
Optionally, the medical image processing apparatus 500 further includes:
the preprocessing module is used for respectively carrying out normalization processing and filtering processing on the medical image picture before the medical image picture is input into a hierarchical network model obtained through pre-training so as to obtain a preprocessed picture;
the training module is specifically configured to input the preprocessed picture into a hierarchical network model obtained through pre-training, and the hierarchical network model calculates the preprocessed picture to output a hierarchical array including C-1 elements.
Optionally, the apparatus further comprises:
and the determining module is used for determining the lesion grade corresponding to the medical image picture according to the grading array.
The apparatus provided in this embodiment can implement each process implemented in the method embodiments shown in fig. 1 to fig. 3, and can achieve the same beneficial effects, and for avoiding repetition, details are not described here again.
According to an embodiment of the present application, an electronic device and a readable storage medium are also provided.
Fig. 7 is a block diagram of an electronic device of a medical image processing method according to an embodiment of the present application. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be examples only, and are not meant to limit implementations of the present application that are described and/or claimed herein.
As shown in fig. 7, the electronic apparatus includes: one or more processors 701, a memory 702, and interfaces for connecting the various components, including a high-speed interface and a low-speed interface. The various components are interconnected using different buses and may be mounted on a common motherboard or in other manners as desired. The processor may process instructions for execution within the electronic device, including instructions stored in or on the memory to display graphical information of a GUI on an external input/output apparatus (such as a display device coupled to the interface). In other embodiments, multiple processors and/or multiple buses may be used, along with multiple memories and multiple memories, as desired. Also, multiple electronic devices may be connected, with each device providing portions of the necessary operations (e.g., as a server array, a group of blade servers, or a multi-processor system). In fig. 7, one processor 701 is taken as an example.
The memory 702 is a non-transitory computer readable storage medium as provided herein. The memory stores instructions executable by at least one processor to cause the at least one processor to perform the medical image processing method provided by the present application. The non-transitory computer-readable storage medium of the present application stores computer instructions for causing a computer to execute the medical image processing method provided by the present application.
The memory 702 serves as a non-transitory computer-readable storage medium, and can be used for storing non-transitory software programs, non-transitory computer-executable programs, and modules, such as program instructions/modules (for example, the acquisition module 501 and the calculation module 502 shown in fig. 5) corresponding to the medical image processing method in the embodiment of the present application. The processor 701 executes various functional applications of the server and data processing by executing the non-transitory software programs, instructions and modules stored in the memory 702, so as to implement the medical image processing method in the above method embodiment.
The memory 702 may include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function; the storage data area may store data created according to use of the electronic device of the medical image processing method, and the like. Further, the memory 702 may include high speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid state storage device. In some embodiments, the memory 702 may optionally include a memory remotely disposed from the processor 701, and these remote memories may be connected to the electronic device of the medical image processing method through a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The electronic device of the medical image processing method may further include: an input device 703 and an output device 704. The processor 701, the memory 702, the input device 703 and the output device 704 may be connected by a bus or other means, and fig. 7 illustrates an example of a connection by a bus.
The input device 703 may receive input numeric or character information and generate key signal inputs related to user settings and function control of the electronic apparatus of the medical image processing method, such as a touch screen, a keypad, a mouse, a track pad, a touch pad, a pointing stick, one or more mouse buttons, a track ball, a joystick, or other input devices. The output devices 704 may include a display device, auxiliary lighting devices (e.g., LEDs), and tactile feedback devices (e.g., vibrating motors), among others. The display device may include, but is not limited to, a Liquid Crystal Display (LCD), a Light Emitting Diode (LED) display, and a plasma display. In some implementations, the display device can be a touch screen.
Various implementations of the systems and techniques described here can be realized in digital electronic circuitry, integrated circuitry, application specific ASICs (application specific integrated circuits), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs that are executable and/or interpretable on a programmable system including at least one programmable processor, which may be special or general purpose, receiving data and instructions from, and transmitting data and instructions to, a storage system, at least one input device, and at least one output device.
These computer programs (also known as programs, software applications, or code) include machine instructions for a programmable processor, and may be implemented using high-level procedural and/or object-oriented programming languages, and/or assembly/machine languages. As used herein, the terms "machine-readable medium" and "computer-readable medium" refer to any computer program product, apparatus, and/or device (e.g., magnetic discs, optical disks, memory, Programmable Logic Devices (PLDs)) used to provide machine instructions and/or data to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal. The term "machine-readable signal" refers to any signal used to provide machine instructions and/or data to a programmable processor.
To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic, speech, or tactile input.
The systems and techniques described here can be implemented in a computing system that includes a back-end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back-end, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), Wide Area Networks (WANs), and the Internet.
The computer system may include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other.
According to the technical scheme of the embodiment of the application, the probability that the lesion grade corresponding to the medical image picture is larger than or equal to each lesion grade is calculated respectively, so that even if errors exist in part of the calculated probability values of each probability value, the lesion grade corresponding to the medical image picture can be predicted by combining other probability values, the problem of reliability of a medical image picture processing result is solved, and the reliability of the prediction result is improved.
It should be understood that various forms of the flows shown above may be used, with steps reordered, added, or deleted. For example, the steps described in the present application may be executed in parallel, sequentially, or in different orders, and the present invention is not limited thereto as long as the desired results of the technical solutions disclosed in the present application can be achieved.
The above-described embodiments should not be construed as limiting the scope of the present application. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may be made in accordance with design requirements and other factors. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present application shall be included in the protection scope of the present application.
Claims (10)
1. A method of medical image processing, comprising:
acquiring a medical image picture;
inputting the medical image picture into a hierarchical network model obtained by pre-training, calculating the medical image picture by the hierarchical network model, and outputting a hierarchical array comprising C-1 elements, wherein the medical image picture corresponds to C lesion levels, the C lesion levels comprise 1 lesion level without lesions and C-1 lesion level with lesions, the kth element of the hierarchical array is the probability that the lesion level corresponding to the medical image picture is greater than or equal to the kth level, and k is greater than or equal to 1 and is less than or equal to C-1.
2. The method of claim 1, wherein prior to acquiring the medical image picture, the method further comprises:
constructing a convolutional neural network model;
and inputting training data to the convolutional neural network model, and training the convolutional neural network model according to the training data to obtain the hierarchical network model.
3. The method of claim 2, wherein the training data comprises a sample image picture and a target array corresponding to the sample image picture, the target array comprising C-1 elements;
when the lesion grade corresponding to the sample image picture is the lesion grade without lesions, the values of C-1 elements in the target array are all 0;
and when the lesion grade corresponding to the sample image picture is the kth grade, the values of the first k elements of the target array are all 1, and the values of the (k + 1) th to the (C-1) th elements of the target array are all 0.
4. The method of claim 1, wherein before inputting the medical image picture into a pre-trained hierarchical network model, the method further comprises:
respectively carrying out normalization processing and filtering processing on the medical image picture to obtain a preprocessed picture;
the step of inputting the medical image picture into a hierarchical network model obtained by pre-training, and calculating the medical image picture by the hierarchical network model to output a hierarchical array comprising C-1 elements comprises:
inputting the preprocessed picture into a hierarchical network model obtained by pre-training, and calculating the preprocessed picture by the hierarchical network model to output a hierarchical array comprising C-1 elements.
5. The method according to claim 1, wherein after the medical image picture is input into a pre-trained hierarchical network model, and the medical image picture is computed by the hierarchical network model to output a hierarchical array comprising C-1 elements, the method further comprises:
and determining the lesion grade corresponding to the medical image picture according to the grading array.
6. A medical image processing apparatus, comprising:
the acquisition module is used for acquiring a medical image picture;
the calculation module is used for inputting the medical image picture into a hierarchical network model obtained by pre-training, calculating the medical image picture by the hierarchical network model and outputting a hierarchical array comprising C-1 elements, wherein the medical image picture corresponds to C lesion levels, the C lesion levels comprise 1 lesion level without lesion and C-1 lesion level with lesion, the kth element of the hierarchical array is the probability that the lesion level corresponding to the medical image picture is greater than or equal to the kth level, and k is greater than or equal to 1 and is less than or equal to C-1.
7. The apparatus of claim 6, further comprising:
the construction module is used for constructing a convolutional neural network model before the medical image picture is acquired;
and the training module is used for inputting training data to the convolutional neural network model and training the convolutional neural network model according to the training data to obtain the hierarchical network model.
8. The apparatus of claim 7, wherein the training data comprises a sample image picture and a target array corresponding to the sample image picture, the target array comprising C-1 elements;
when the lesion grade corresponding to the sample image picture is the lesion grade without lesions, the values of C-1 elements in the target array are all 0;
and when the lesion grade corresponding to the sample image picture is the kth grade, the values of the first k elements of the target array are all 1, and the values of the (k + 1) th to the (C-1) th elements of the target array are all 0.
9. An electronic device, comprising:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-5.
10. A non-transitory computer readable storage medium having stored thereon computer instructions for causing the computer to perform the method of any one of claims 1-5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010076247.7A CN111275121B (en) | 2020-01-23 | 2020-01-23 | Medical image processing method and device and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010076247.7A CN111275121B (en) | 2020-01-23 | 2020-01-23 | Medical image processing method and device and electronic equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111275121A true CN111275121A (en) | 2020-06-12 |
| CN111275121B CN111275121B (en) | 2023-07-18 |
Family
ID=71003345
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010076247.7A Active CN111275121B (en) | 2020-01-23 | 2020-01-23 | Medical image processing method and device and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111275121B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111754486A (en) * | 2020-06-24 | 2020-10-09 | 北京百度网讯科技有限公司 | Image processing method, image processing device, electronic equipment and storage medium |
| CN112016634A (en) * | 2020-09-30 | 2020-12-01 | 北京百度网讯科技有限公司 | Medical image recognition method, device, equipment and storage medium |
| WO2021120589A1 (en) * | 2020-06-17 | 2021-06-24 | 平安科技（深圳）有限公司 | Method and apparatus for abnormal image filtering for use on 3d images, device, and storage medium |
Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120155726A1 (en) * | 2009-08-24 | 2012-06-21 | Huiqi Li | method and system of determining a grade of nuclear cataract |
| US20170103611A1 (en) * | 2015-10-12 | 2017-04-13 | Diamond Game Enterprises | System and method for using conditional probabilities to enhance gaming payouts |
| CN106909778A (en) * | 2017-02-09 | 2017-06-30 | 北京市计算中心 | A kind of Multimodal medical image recognition methods and device based on deep learning |
| CN106971198A (en) * | 2017-03-03 | 2017-07-21 | 北京市计算中心 | A kind of pneumoconiosis grade decision method and system based on deep learning |
| CN107145904A (en) * | 2017-04-28 | 2017-09-08 | 北京小米移动软件有限公司 | Determination method, device and the storage medium of image category |
| CN107832838A (en) * | 2017-11-28 | 2018-03-23 | 北京羽医甘蓝信息技术有限公司 | The method and apparatus for evaluating cell smear sample satisfaction |
| US20180309990A1 (en) * | 2015-10-13 | 2018-10-25 | Samsung Electronics Co., Ltd. | Method and device for encoding or decoding image |
| WO2018201647A1 (en) * | 2017-05-05 | 2018-11-08 | 平安科技（深圳）有限公司 | Method for detecting retinopathy degree level, device and storage medium |
| CN109215027A (en) * | 2018-10-11 | 2019-01-15 | 平安科技（深圳）有限公司 | A kind of car damage identification method neural network based, server and medium |
| CN109308564A (en) * | 2018-08-01 | 2019-02-05 | 平安科技（深圳）有限公司 | The recognition methods of crowd's performance ratings, device, storage medium and computer equipment |
| CN109686444A (en) * | 2018-12-27 | 2019-04-26 | 上海联影智能医疗科技有限公司 | System and method for medical image classification |
| CN110047593A (en) * | 2019-04-12 | 2019-07-23 | 平安科技（深圳）有限公司 | Disease popularity season grade determination method, apparatus, equipment and readable storage medium storing program for executing |
| CN110136103A (en) * | 2019-04-24 | 2019-08-16 | 平安科技（深圳）有限公司 | Medical image means of interpretation, device, computer equipment and storage medium |
| US10430946B1 (en) * | 2019-03-14 | 2019-10-01 | Inception Institute of Artificial Intelligence, Ltd. | Medical image segmentation and severity grading using neural network architectures with semi-supervised learning techniques |
| CN110544014A (en) * | 2019-08-06 | 2019-12-06 | 中国航天标准化研究所 | Aerospace safety risk classification and deep cause analysis method and model |
-
2020
- 2020-01-23 CN CN202010076247.7A patent/CN111275121B/en active Active
Patent Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120155726A1 (en) * | 2009-08-24 | 2012-06-21 | Huiqi Li | method and system of determining a grade of nuclear cataract |
| US20170103611A1 (en) * | 2015-10-12 | 2017-04-13 | Diamond Game Enterprises | System and method for using conditional probabilities to enhance gaming payouts |
| US20180309990A1 (en) * | 2015-10-13 | 2018-10-25 | Samsung Electronics Co., Ltd. | Method and device for encoding or decoding image |
| CN106909778A (en) * | 2017-02-09 | 2017-06-30 | 北京市计算中心 | A kind of Multimodal medical image recognition methods and device based on deep learning |
| CN106971198A (en) * | 2017-03-03 | 2017-07-21 | 北京市计算中心 | A kind of pneumoconiosis grade decision method and system based on deep learning |
| CN107145904A (en) * | 2017-04-28 | 2017-09-08 | 北京小米移动软件有限公司 | Determination method, device and the storage medium of image category |
| WO2018201647A1 (en) * | 2017-05-05 | 2018-11-08 | 平安科技（深圳）有限公司 | Method for detecting retinopathy degree level, device and storage medium |
| CN107832838A (en) * | 2017-11-28 | 2018-03-23 | 北京羽医甘蓝信息技术有限公司 | The method and apparatus for evaluating cell smear sample satisfaction |
| CN109308564A (en) * | 2018-08-01 | 2019-02-05 | 平安科技（深圳）有限公司 | The recognition methods of crowd's performance ratings, device, storage medium and computer equipment |
| CN109215027A (en) * | 2018-10-11 | 2019-01-15 | 平安科技（深圳）有限公司 | A kind of car damage identification method neural network based, server and medium |
| CN109686444A (en) * | 2018-12-27 | 2019-04-26 | 上海联影智能医疗科技有限公司 | System and method for medical image classification |
| US10430946B1 (en) * | 2019-03-14 | 2019-10-01 | Inception Institute of Artificial Intelligence, Ltd. | Medical image segmentation and severity grading using neural network architectures with semi-supervised learning techniques |
| CN110047593A (en) * | 2019-04-12 | 2019-07-23 | 平安科技（深圳）有限公司 | Disease popularity season grade determination method, apparatus, equipment and readable storage medium storing program for executing |
| CN110136103A (en) * | 2019-04-24 | 2019-08-16 | 平安科技（深圳）有限公司 | Medical image means of interpretation, device, computer equipment and storage medium |
| CN110544014A (en) * | 2019-08-06 | 2019-12-06 | 中国航天标准化研究所 | Aerospace safety risk classification and deep cause analysis method and model |
Non-Patent Citations (3)
| Title |
|---|
| GANG LI等: ""Noise image segmentation model with local intensity difference"", 《JOURNAL OF COMPUTER APPLICATION》 * |
| 吴树刚等: ""岩质路堑边坡概率分级模型研究"", 《山西建筑》 * |
| 陈涛等: ""计算机视觉技术在农产品品质检测中的应用"", 《安徽农学通报》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021120589A1 (en) * | 2020-06-17 | 2021-06-24 | 平安科技（深圳）有限公司 | Method and apparatus for abnormal image filtering for use on 3d images, device, and storage medium |
| CN111754486A (en) * | 2020-06-24 | 2020-10-09 | 北京百度网讯科技有限公司 | Image processing method, image processing device, electronic equipment and storage medium |
| US11379980B2 (en) | 2020-06-24 | 2022-07-05 | Beijing Baidu Netcom Science Technology Co., Ltd. | Image processing method, apparatus, electronic device and storage medium |
| CN111754486B (en) * | 2020-06-24 | 2023-08-15 | 北京百度网讯科技有限公司 | Image processing method, device, electronic equipment and storage medium |
| CN112016634A (en) * | 2020-09-30 | 2020-12-01 | 北京百度网讯科技有限公司 | Medical image recognition method, device, equipment and storage medium |
| CN112016634B (en) * | 2020-09-30 | 2023-07-28 | 北京百度网讯科技有限公司 | Medical image recognition method, device, equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111275121B (en) | 2023-07-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111710412B (en) | Diagnostic result verification method and device and electronic equipment | |
| CN111275121A (en) | Medical image processing method and device and electronic equipment | |
| CN112329964B (en) | Method, device, equipment and storage medium for pushing information | |
| CN111967538A (en) | Feature fusion method, device and equipment applied to small target detection and storage medium | |
| US20210117484A1 (en) | Webpage template generation | |
| US20220027854A1 (en) | Data processing method and apparatus, electronic device and storage medium | |
| CN111832298A (en) | Quality inspection method, device and equipment for medical records and storage medium | |
| CN111738014A (en) | Drug classification method, device, equipment and storage medium | |
| CN112329453B (en) | Method, device, equipment and storage medium for generating sample chapter | |
| CN111916203B (en) | Health detection method, apparatus, electronic device and storage medium | |
| CN111695680B (en) | Score prediction method, score prediction model training method and device and electronic equipment | |
| CN112732940A (en) | Model-based medical knowledge graph reasoning method, device, equipment and medium | |
| CN112541924A (en) | Fundus image generation method, device, equipment and storage medium | |
| Azeez et al. | Secondary triage classification using an ensemble random forest technique | |
| CN111754486B (en) | Image processing method, device, electronic equipment and storage medium | |
| CN113838573B (en) | Clinical assistant decision-making diagnosis self-learning method, device, equipment and storage medium | |
| CN111767946A (en) | Medical image hierarchical model training and prediction method, device, equipment and medium | |
| CN111785340A (en) | Medical data processing method, device, equipment and storage medium | |
| CN111524579B (en) | Lung function curve detection method, device, equipment and storage medium | |
| CN110706772B (en) | Ordering method and device, electronic equipment and storage medium | |
| JP7346419B2 (en) | Learning and applying contextual similarities between entities | |
| CN112733879A (en) | Model distillation method and device for different scenes | |
| CN112579875A (en) | Method, device, equipment and medium for generating release information title | |
| CN115153517B (en) | Testing method, device, equipment and storage medium for timing, standing and walking test | |
| CN116092616B (en) | Medical data transmission method, device, equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20210416 Address after: 100000 Room 220, 2nd Floor, Building 4, No. 1, Shangdi East Road, Haidian District, Beijing Applicant after: Beijing Confucius Health Technology Co.,Ltd. Address before: 100085 Beijing, Haidian District, No. ten on the ground floor, No. 10 Baidu building, layer 2 Applicant before: BEIJING BAIDU NETCOM SCIENCE AND TECHNOLOGY Co.,Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |